Hi,

Sorry if the subject is confusing.

I assume that his is a similar setup to having a motor driving something mechanical and then using a '3rd party' measurement system to feedback to the IP-A. Linear scales is a good example of this.

In my case I have a DC motor, that takes -10 to +10V input. All good, IP-A, sorted. The DC motor then has Tach feedback to its own Drive.

Then, on the ballscrew pulley (not the motor), is a suitable encoder (5V TTL) that can give the IP-A unit feedback.

The pulleys are not the same size, therefore the ratio is not the same as the motor speed. The ratio is however fixed. The Encoder does 2000 PPR measurement.

Is it possible for the IP-A unit to be used in this situation? If so, how? Where do I configure the calibration or do I mess with the steps per mm count to make the maths work?

Many thanks.

Chaz,

The config you describe should work fine. I have a similar config in that servo amps receive only tach feedback from servos. On my config encoders are on the servo motor shafts however I see no reason why your config should not work. Just compute the steps/inch or steps/mm correctly according to the location of your encoders. Since you have 2000L encoders they provide 4 edges per pulse or 8k/rev edges. If encoders are on ball screws directly you simply divide 8k/pitch to arrive at the correct scaling. On my machine I have .2"/rev pitch so I would calculate 8k/.2" to get the correct scale. The scale is entered In Mach under Config-->Motor Tuning.

In theory it may be more accurate to have encoders directly on ball screws if, like my machine, screws are driven by belts. It would seem that belts stretching could affect the accuracy of my machine. If however the encoder is fixed to the screw then belts could not affect accuracy. Does this make sense?

Good Luck!
Mike

Chaz,

The config you describe should work fine. I have a similar config in that servo amps receive only tach feedback from servos. On my config encoders are on the servo motor shafts however I see no reason why your config should not work. Just compute the steps/inch or steps/mm correctly according to the location of your encoders. Since you have 2000L encoders they provide 4 edges per pulse or 8k/rev edges. If encoders are on ball screws directly you simply divide 8k/pitch to arrive at the correct scaling. On my machine I have .2"/rev pitch so I would calculate 8k/.2" to get the correct scale. The scale is entered In Mach under Config-->Motor Tuning.

In theory it may be more accurate to have encoders directly on ball screws if, like my machine, screws are driven by belts. It would seem that belts stretching could affect the accuracy of my machine. If however the encoder is fixed to the screw then belts could not affect accuracy. Does this make sense?

Good Luck!
Mike

Hi, yes, perfectly. I thought about it last night again, it must work. And as its on the drive shaft, as you say, more accurate. It doesnt care what the motor does, it cares where the leadscrew 'is'.

Thanks, will confirm later today hopefully once I get this running.

So, this is working although we had issues on using Enc Channel 0. I checked wiring, I checked everything. X kept coming up with an error although it would move, the CSMIO Plugin highlights a problem. Ill upload a vid. After numerous attempts to fix, mechanical and electrical, I moved it to Enc Channel 3, problem gone. Any ideas? This is whilst the Y Axis, wired exactly the same way, works without issue. Just upload a vid of this to utoob, will link when done.

Vid is not very clear, hoping that its utoob busy processing. If you watch the X values, they increment up and down but flash 255 in red too. What does this mean?

https://www.youtube.com/watch?v=h4sDZypSXXM

So have a response from CS Labs. The red numbers are a count of errors. Suggested that I check shielding etc.

Good to know what the red means

Regarding using scales or encoders mounted remotely from the motors, you will only have issues if there is backlash between motor and encoder/scale, in that case you may have tuning issues.

The most accurate location for a scale is on the axis itself as it will be the true position.
On the shaft of the ballscrew you will still be on the wrong side of backlash, if there is any in the ballnut, so it will not really be any more accurate than having the encoder on the motor.
Hood

Thanks. Started to do the retrofit properly yesterday. As of last night managed to get everything working. Had a few small issues, mostly own mistakes.

This is a really interesting machine, it was built in the 90s for we think Datron to buy the rights too. Granite base, really small, air bearing spindle etc.

Need to do a bit more wiring and see if we can reuse the very expensive Euchner MPG that comes from the stuff we have removed.

Some pics here, the wiring is not final, will be tidied when it goes into the case.

https://www.flickr.com/photos/79889472@N06/albums/72157663462959562

This is the machine itself.

http://www.mycncuk.com/threads/9377-Any-experience-or-software-for-Control-Techniques-FNC4-Wha-do-I-replace-it-with

Nice looking mill, are the slides air bearing as well or are there profile rails in there somewhere?
Hood